home-assistant/core · error · HomeAssistantError

AI Task entity {entity_id} does not support generating image

Error message

AI Task entity {entity_id} does not support generating images

What it means

HomeAssistantError from ai_task.async_generate_image when the entity exists but supported_features lacks AITaskEntityFeature.GENERATE_IMAGE. Only agents whose backend can produce images (and whose integration implements internal_async_generate_image) advertise this flag; text-only agents are rejected here before the chat session is even opened.

Source

Thrown at homeassistant/components/ai_task/task.py:182

    *,
    task_name: str,
    entity_id: str | None = None,
    instructions: str,
    attachments: list[dict] | None = None,
) -> ServiceResponse:
    """Run an image generation task in the AI Task integration."""
    if entity_id is None:
        entity_id = hass.data[DATA_PREFERENCES].gen_image_entity_id

    if entity_id is None:
        raise HomeAssistantError("No entity_id provided and no preferred entity set")

    entity = hass.data[DATA_COMPONENT].get_entity(entity_id)
    if entity is None:
        raise HomeAssistantError(f"AI Task entity {entity_id} not found")

    if AITaskEntityFeature.GENERATE_IMAGE not in entity.supported_features:
        raise HomeAssistantError(
            f"AI Task entity {entity_id} does not support generating images"
        )

    if (
        attachments
        and AITaskEntityFeature.SUPPORT_ATTACHMENTS not in entity.supported_features
    ):
        raise HomeAssistantError(
            f"AI Task entity {entity_id} does not support attachments"
        )

    with async_get_chat_session(hass) as session:
        resolved_attachments = await _resolve_attachments(hass, session, attachments)

        task_result = await entity.internal_async_generate_image(
            session,
            GenImageTask(
                name=task_name,

View on GitHub (pinned to 58a3fdb3ea)

Solutions

Target an image-capable agent: pass the entity_id of an integration whose models generate images (current OpenAI image models via the OpenAI conversation integration, Google Gemini image models, etc.).
Update the provider integration to a version implementing AI Task image generation.
Custom agents: implement internal_async_generate_image and set AITaskEntityFeature.GENERATE_IMAGE.
For text-only agents, use ai_task.generate_data instead.

Defensive patterns

Strategy: validation

Validate before calling

def supports_gen_image(entity) -> bool:
    """Entity advertises image-generation capability."""
    return AITaskEntityFeature.GENERATE_IMAGE in entity.supported_features

Type guard

def is_image_capable_ai_task_entity(entity) -> bool:
    """Narrowing guard: registered and GENERATE_IMAGE-capable."""
    return (
        hasattr(entity, "supported_features")
        and AITaskEntityFeature.GENERATE_IMAGE in entity.supported_features
    )

Try / catch

from homeassistant.exceptions import HomeAssistantError

try:
    resp = await async_generate_image(hass, entity_id=eid, task_name="img", instructions="...")
except HomeAssistantError as err:
    if "does not support generating images" in str(err):
        # switch to an image-capable agent id and retry
        raise
    raise

Prevention

Point generate_image only at agents with image models behind them.
Update provider integrations so the image feature flag is implemented.
Custom agents: implement internal_async_generate_image and set the flag.

When it happens

Trigger: Calling ai_task.generate_image against a text-only conversation agent — entity resolves fine, but the GENERATE_IMAGE bit is absent.

Common situations: Default preference left on a plain chat model; local LLM setups (Ollama etc.) without image output; provider integrations that have not implemented the image API yet.

Related errors

AI-assisted analysis of home-assistant/core@58a3fdb3ea (2026-08-14). Data as JSON: /api/errors/47120f3a440d9700. Report an issue: GitHub.